GASTROENTEROLOGY 1991;101:167-174

Positive Serum Antibody and Negative

Tissue Staining for Helicobacfer pylori in
Subjects With Atrophic Body Gastritis

W. E. KARNES, Jr., I. M. SAMLOFF, M. SIURALA, M. KEKKI,

P. SIPPONEN, S. W. R. KIM, and J. H. WALSH
Center for Ulcer Research and Education, Veterans Administration Medical Center, UCLA, Los
Angeles, and Veterans Administration Medical Center, Sepulveda, California: Second
Department of Medicine, University of Helsinki, Helsinki, and Jorvi Hospital, Espoo, Finland

Helicobacterpylori is rarely found in gastric biopsy
specimens from individuals with atrophic gastritis
of the body mucosa. To determine if subjects with
atrophic body gastritis have evidence of previous
infection with H. pylori, immunoglobulin G anti-
body to H. pylori was measured by enzyme-linked
immunosorbent assay in sera of 399 Finnish sub-
jects. In 124 subjects, multiple biopsy specimens
from body and antrum had been evaluated for the
presence of H. pylori by Giemsa staining. Antibody
correlated well with H. pylori staining except in the
subgroup with atrophic body gastritis, in whom the
prevalence of seropositivity (86%) was significantly
greater than the prevalence of positive staining
(33%) (P < 0.001). Twenty-five subjects had positive
antibody and negative staining. This group had a
significantly higher prevalence of atrophic body
gastritis (BO%), lower maximal acid output, lower
serum pepsinogen I levels, and higher serum gastrin
concentrations than did seropositive subjects with
H. pylori. These data suggest that most patients with
atrophic body gastritis, despite having a low inci-
dence of current overt infection, have been infected
with H. pylori at some point in their lives.
suggesting that H. pylori-associated superficial gastri-
tis may progress over time into atrophic gastritis.
In a longitudinal study examining the natural his-
tory of gastritis over three decades in a large Finnish
population, Ihamaki et al. (8) found that superficial
gastritis affecting the antrum tended to regress with
time, whereas superficial gastritis affecting the body
tended to progress to atrophic body gastritis. Al-
though their study predated the discovery of H. pylori,
it is tempting to speculate that the cases of superficial
gastritis that progressed to atrophic body gastritis
were caused by H. pylori. However, because the
prevalence of H, pylori infection in subjects with
atrophic body gastritis is quite low (3), the progres-
sion of H. pylori-associated superficial gastritis to
atrophic body gastritis must be accompanied by disap-
pearance of H. pylori.
To determine if atrophic body gastritis is associated
with evidence of previous infection with H. pylori, we
measured H. pylori antibody in sera from 399 normal
Finnish volunteers in whom histology of the gastric
body and antrum were known. The results of Giemsa
staining of gastric biopsies for H. pylori had been
previously determined in 124 subjects. Additional
data collected in all 399 subjects included age, sex,
maximal acid output, serum pepsinogen I (PGI), PGII,
and fasting plasma gastrin concentration. These data

H elicobacter pylori is the probable cause of most

cases of superficial gastritis of the antrum and
body (l-6). The prevalence of H. pylori colonization
allowed us to further characterize the group of sub-
jects who tested positive for antibody but negative for
Giemsa stains.
of the gastric mucosa increases with age, reaching
50% by age 50 in developed countries and nearly 80%
by age 30 in underdeveloped parts of the world (7). In
Abbreviations used in this paper: MAO, maximal acid output;
a given population, the prevalence of atrophic gastri-
PGI, pepsinogen I.
tis among older age groups correlates well with the o 1991 by the American Gastroenterological Association
prevalence of H. pylori in younger age groups (l), 9916~5065/91/$X00
168 KARNES ET AL.
Methods
Subjects
The present study included 399 stored sera collected
from members of a control group originally selected from a
large Finnish population for comparison with subjects with
gastric carcinoma (9). They included 67 control subjects
computer matched to 67 probands with gastric carcinoma.
The remaining 332 sera were from first-degree relatives of
the control subjects. Matching was by age, sex, place of
birth, and place of residence. A history of duodenal ulcer
was present in 3 subjects, and 1 had a history of gastric
ulcer. Most had no upper abdominal complaints. Among
the original data collected in all subjects were age, sex,
pentagastrin-stimulated acid secretion and fasting serum
concentrations of gastrin (lo],PGI and PGII (1I]concentra-
tions, and histological interpretation of gastric biopsy speci-
mens. Sera from subjects with atrophic body gastritis
[determined as described below) were evaluated for vitamin
B,, levels and antibody to intrinsic factor. Also included
were Giemsa staining results for H. pylori in 124 subjects as
described below. Serum and gastric biopsy specimens were
obtained on the same day.
His tology
Gastric biopsy specimens were obtained endoscopi-
tally under direct vision in all subjects as described previ-
ously (3). Four or more specimens were taken from the
middle part of the antrum at least 2 cm from the pylorus,
and six or more specimens were obtained from the anterior
wall and greater curvature of the midbody. Biopsy speci-
mens were fixed overnight in buffered formalin, embedded
in paraffin, sectioned, and stained with Alcian blue, peri-
odic acid-Schiff, and van Gieson for histology.
Histology of the antrum and body was classified accord-
ing to a modification of previously described criteria (12) as
normal mucosa, 1;mild, moderate, and severe superficial
gastritis, 2, 3, and 4, respectively, based on the degree of
inflammatory infiltration in the setting of normal gastric
glands; and mild, moderate, and severe atrophic gastritis, 5,
6,and 7, respectively, based on degree of glandular atrophy,
irrespective of inflammatory infiltration. With few excep-
tions, histological grades in adjacent biopsy samples from
the same patient were similar. The biopsy specimen with
the highest grade was assigned when histological findings
in adjacent biopsy specimens were dissimilar.
Stainingfor H. pylori
Of the 399 subjects, 124 had adequate tissue blocks
available for detection of H. pylori using a modified Giemsa
technique (13). The selection of these subjects has been
previously described (3). On average, 11 biopsy specimens
were stained from each patient. Availability of adequate
material allowed for staining of both body and antral
biopsies in 110 subjects, antrum alone in 9 subjects, and
body alone in 5 subjects. Subjects with H. pylori present in
any section from the antrum, body, or both were considered
to be positive.
GASTROENTEROLOGYVol.1o1,No.1
Enzyme-Linked Immunosorbent Assay
Antibody to H. pylori was measured by enzyme-
linked immunosorbent assay (ELISA) in all 399 sera. Be-
cause of its reported high specificity (la), partially purified
H. pylori urease was selected as antigen for this study. Using
this antigen, our ELISA has a specificity of 100% and a
sensitivity of 93%, using [‘C]urea breath test results as the
“gold standard” (I 5).
As previously described (15), an H. pylori isolate was
subjected to French Press (AMINCO, Urbane, IL) disruption
followed by centrifugation. The supernatant was separated
by Sephacryl S-400 column chromatography, and the frac-
tions with peak urease activity were pooled and diluted to
10 &mL in phosphate-buffered saline (PBS) for use as
antigen. Test sera and 16 negative control sera were assayed
at a dilution of 1:lOOOin duplicate. Goat anti-human IgG
conjugated with horse-radish peroxidase was used as sec-
ond antibody. Reaction with substrate (o-phenylenedi-
amine and hydrogen peroxide] was stopped after 36 min-
utes with 10% H,SO,, and the absorbance at 492 nm was
measured using an ELISA reader (Titertek MkII; Flow
Laboratories).
H. pylori seropositivity was defined by an absorbance > 2
standard deviations above the mean absorbance of 16 sera
obtained from subjects with negative [‘%]urea breath test
results.
Campylobacter jejuni Enzyme-Linked
lmmunosorbent Assay
Campylobacter jejuni was kindly provided by Dr.
Perez-Perez (Division of Infectious Disease, Vanderbilt Uni-
versity, Nashville, TN). C. jejuni was grown on Campylobac-
ter-selective blood agar (BBL Microbiology Systems, Cock-
eysville, MD) in microaerobic atmosphere at 37°C. Colonies
were scraped free of the agar, washed twice with 0.2mol/L
PBS (pH 7.4), and subjected to French Press disruption with
a cell pressure of 12,000lb/in* (Aminco, Urbana, II). French
Press extract was clarified by centrifugation at 569Og for 36
minutes, and protein concentration was measured by the
BioRad microassay procedure (Bio-Rad Chemical Division,
Richmond, CA). The ELISA plates were coated with 10
p&nL of the C. jejuni antigen for use in ELISA by the same
method that was used for H. pylori ELISA.
Statistics
x2 and stepwise logistic regression analysis, and
Student’s t test were used for statistical calculations. Devia-
tions were expressed as standard error (SE). Significance
was defined as P < 0.05.
Results
Analysis of 399 Subjects
Table 1 compares seropositive and seronega-
tive subjects with respect to mean age, sex, serum
PGs, serum gastrin, maximal acid output (MAO), and
July 1991 H. PYLORI ANTIBODY AND STAINING IN ATROPHIC BODY GASTRITIS 169

Table 1. Comparison of Measured Parameters Between and tended to decrease with increasing severity of
Subjects With Positive Serum Antibody and atrophic gastritis of the body, although the trend
Negative Serum Antibody to H. pylori downward was not significant by x2 trend analysis.
H. pylori antibody Seropositivity remained high in subjects with atro-
phic antral gastritis. Of the 210 subjects with positive
Positive Negative
n = 210 n = 189 P
H. pylori serology, 207 (99%) had some degree of
gastritis affecting either the body or antrum. The
Sex (% male) 51 46 NS
prevalence of seropositivity was 13% (20 of 153) in
Age (yr) 55.2 (1.0) 41.6 (1.2) < 0.001
PGI (ng/mL) 79.3 (2.9) 65.5 (2.3) < 0.001”
subjects with normal antral histology, and 9% (14 of
PGII (ng/mL) 23.3 (0.8) 12.7 (0.6) < 0.001 152) in subjects with normal body histology. Of the
PGI-PGII ratio 3.7 (0.1) 6.0 (0.2) < 0.001” 132 subjects who had normal histology of both the
Gastrin (pg/mL) 68.4 (4.7) 44.6 (1.1) < 0.001 antrum and body only 4 (3.3%) had positive serology
MAO (mmolih) 20.7 (1.1) 27.2 (1.0) < 0.001
for H. pylori.
Antral grade (1-7) 3.5 (0.1) 1.8 (0.1) < 0.001”
Body grade (l-7) 3.3 (0.1) 1.7 (0.1) < 0.001”

NOTE. Results are expressed as mean -CSE.
“Significant distinguishing variables (P < 0.05) by stepwise logis-
tic regression analysis.
degree of antral and body gastritis. Seropositive sub-
jects were significantly older, had lower maximal acid
outputs, greater severity of antral and body gastritis,
higher serum concentrations of gastrin, PGI, and PGII,
and a significantly lower PGI-PGII ratio. There was no
gender predilection in the seropositive group. Step-
wise logistic regression analysis showed that the
significant (P < 0.05) predictors of seropositivity were
degree of antral gastritis, degree of body gastritis, and
the serological markers of gastritis: PGI-PGII ratio and
PGI levels. The remaining parameters, PGII, MAO,
gastrin, age, and sex, were not significant predictors
of H. pylori antibody status.
Seropositivity increased from 19% in the 14-20-
year-old age group to 74% in the 71-BO-year-old age
group (Figure 1). Positive serology was also related to
degree of gastritis, increasing abruptly with any de-
gree of superficial gastritis of the body or antrum
(Figure 2). Seropositivity rate was highest in subjects
with severe superficial gastritis of the body (grade 4)
N whelming predominance
Figure 1. Prevalence of H. pylori with increasing age. Numbers of
subjects within each age group is shown at top.
Analysis of 124 Subjects Evaluated for H.
pylori by Giemsa Stain
Although the group of subjects who were eval-
uated by Giemsa staining were selected on the basis of
availability of tissue blocks, they differed significantly
from those whose biopsy specimens were not stained
by several measured parameters (Table 2). Compared
with the group who were not evaluated for H. pylori
by Giemsa staining, those who had staining were
significantly older, had a significantly higher preva-
lence of H. pylori antibody, lower mean MAO, higher
mean PGI concentration, and higher mean histologi-
cal scores of the antrum and body.
Among the 124 subjects in whom biopsy specimens
were evaluated by Giemsa stain, H. pylori was found
in 67 (54%). Compared with subjects with negative
staining results, the group with positive staining was
older, contained a higher percentage of women, dem-
onstrated more severe antral gastritis, and had signifi-
cantly higher concentrations of PGI and PGII II (Table
3).
Table 4 shows subjects distributed within grades of
body and antral gastritis according to H. pylori stain-
ing and antibody status. The group with negative
staining and negative antibody contained an over-
of subjects with normal
antral and body histology. Among those with positive
staining and positive antibody, 80% had superficial
body gastritis, accompanied by equal proportions of
superficial and atrophic antral gastritis. In contrast, of
those with negative staining and positive antibody
(“false positives”), 80% had atrophic body gastritis
accompanied by equal proportions of normal, superfi-
cial, and atrophic antral histologies. The prevalence
of false positives was significantly higher in the group
with atrophic body gastritis (58%, n = 36) than in
groups with superficial body gastritis (9%, n = 57),
normal body histology (O%, n = 31), or any grade of
antral gastritis. Of the 25 subjects with false positive
antibody, 20 (80%) had atrophic gastritis affecting the
body: of them, 6 (30%) had isolated atrophic body
170 KARNES ET AL. GASTROENTEROLOGY Vol. 101, No. 1

152 92 69 24 17 32 13 (N) 153 60 57 38 63 27 1

100, I I I I I I 1

80

60

40
Figure 2. Prevalence of & py-
20 lori seropositivity with in-
creasing grades of body (A)
f-l and antral gastritis [B). Num-
1234567 bers of subjects are shown
NL SUPERFICIAL ATROPHIC
above each category. *The
A only subject in this group was
Grade of Body Gastritis Grade of Antral Gastritis seropositive.

gastritis (type A gastritis) and 8 (40%) accompanying
atrophic antral gastritis (type AB). The remaining 6
(30%) had accompanying superficial antral gastritis.
None of these subjects had antibody to intrinsic factor
or had abnormal vitamin B,, levels.
H. pylori antibody was found in 59 (88%) of sub-
jects with positive staining for H. pylori and in 25
(44%) of those with the negative staining. Table 5
shows that among seropositive subjects, those with
positive staining did not differ significantly from
those with negative staining with respect to age or
gender. However, the two groups differed signifi-
cantly in severity of gastitis and in serological param-
eters of gastritis. The false-positive group (positive
antibody and negative stain) had a higher score for
body gastritis, a lower score for antral gastritis, a
lower mean MAO, a higher level of serum gastrin, a
lower level of serum PGI, and a lower PGI-PGII ratio.
Figure 3 and Table 6 illustrate that the discrepancy
between positive serology and positive staining for H.
pylori is a function of atrophic gastritis of the body
mucosa. In subjects with normal mucosa and in those
with all grades of superficial gastritis, there was a
close correlation between the prevalence of positive
serology and positive staining. By contrast, in the
subgroup with any degree of atrophic gastritis of the
body mucosa, the prevalence of seropositivity was
higher than the prevalence of positive staining. This
relationship was not evident for the antral mucosa.
The secretory and serological parameters of gastric
body function, MAO and serum PGI level, which
decrease with increasing severity of gastritis, paral-
leled the histological findings, as did serum gastrin
level, which is higher in subjects with atrophic
gastritis.
To determine the specificity of the high prevalence
of seropositivity in subjects with atrophic body gastri-
tis, antibody to C. jejuni was measured by ELISA in all
399 sera. Antiserum to C. jejuni is known to cross-
react with some H. pylori antigens, as well as with
antigens expressed by other enteric organisms (16-
19). Figure 4 shows the lack of correlation between
absorbance readings using our purified H. pylori
antigen and crude C. jejuni antigen among all 399
subjects (r = 0.02). Superimposed over these data are
seropositive subjects who had negative staining
(n = 25) and seropositive subjects who had positive
staining (n = 59). No significant difference between

Table 3. Comparison of Measured Parameters Between

Table 2. Comparison of Measured Parameters Between
Subjects With Positive Tissue Stains and Negative
Subjects Selected for Evaluation of Giemsa
Tissue Stains for H. pylori
Staining for H. pylori and Those Who Had
Inadequate Tissue Blocks for Evaluation H. pylori stain

Stained Not stained Positive Negative

n = 124 n= 275 P Parameters n = 67 n = 57 P

Age (yr) 52.8 (1.5) 46.9 (1.0) <O.OOl Sex (% male) 45 67 <0.05
PGI (ng/mL) 64.9 (3.4) 76.3 (2.3) CO.01 Age(yr) 56.6 (1.7) 48.3 (2.5) <O.Ol
PGII (ng/mL) 18.8 (1.1) 18.0 (0.7) NS PGI (ng/mL) 80.6 (4.6) 46.5 (4.0) <0.0001
PGI-PGII ratio 4.4 (0.3) 5.0 (0.1) ‘co.05 PGII (ng/mL) 22.2 (1.4) 14.7 (1.4) <O.OOl
Gastrin (pg/mL) 65.3 (6.1) 53.5 (3.2) NS PGI-PGII ratio 4.1 (0.2) 4.7 (0.5) NS
h4AO (mmollb) 19.9 (1.4) 25.6 (0.9) <O.OOl Gastrin (pglmt) 62.5 (7.5) 68.7 (10.1) NS
Antral grade (1-7) 3.3 (0.2) 2.4 (0.1) <O.OOOl MAO (mmollh) 22.3 (1.8) 17.1 (2.3) NS
Body grade (l-7) 3.4 (0.2) 2.1 (0.1) <0.0001 Antral grade (l-7) 4.2 (0.2) 2.3 (0.2) <0.0001
Positive H. pylori antibody (%) 68 46 <0.0001 Body grade (l-7) 3.3 (0.2) 3.4 (0.3) NS

NOTE. Results are expressed as mean -C SE. NOTE. Results are expressed as mean k SE.
July 1991 H. PYLORZ ANTIBODY AND STAINING IN ATROPHIC BODY GASTRITIS 171

Table 4. Distribution of Subjects Within Grades ofAntra1 lence of positive staining in those with atrophic body
Gastritis and Body Gastritis According to H. pylori gastritis was 33% (12 of 36 subjects), whereas the
Antibody and Giemsa Stain Results prevalence of positive H. pylori antibody was 86% (31
Histological status of the antrum of 36 subjects). Furthermore, subjects with atrophic
body gastritis accounted for 20 of the 25 false positive
Normal; Superficial; Atrophic;
AB ST grade I(%) grades 2-4 (%) grades 5-7 (%) n H. pylori antibody results (positive antibody with
negative staining). These observations are consistent
- - 91 3 6 32
- + 12 50 38 8
with those reported by Fox et al. (2 7). Analysis of their
+ - 24 36 40 25 data showed that 75% (9 of 12) of false positive ELBA
+ -t 3 46 51 59 results occurred in individuals with atrophic gastritis.
Histological status of the body However, their results were not reported separately
for body and antral atrophic gastritis. To our knowl-
Normal; Superficial; Atrophic;
edge, the present study is the first to compare ELISA
AB ST grade 1 (%) grades 2-4 (%) grades 5-7 (%) n
with H. pylori staining of gastric biopsy specimens in
- - 84 3 13 32 individuals with well characterized histology of the
- + 38 50 1 8
0 20 80 25
antral and body mucosa.
+ -
+ + 2 80 19 59 Among seropositive subjects, the only factor that
seemed to differentiate those with negative stain from
AB, H. pylori antibody results: ST, H. pylori stain results.
those with positive stain results was atrophic body
gastritis. Other parameters that were significantly
these groups was found with respect to C. jejuni different between these two groups were PGI, PGI-
ELISA results. PGII ratio, gastrin, and MAO (Table 4). All of these
factors are known to be influenced by the histological
status of the body mucosa. Interestingly, there was no
Discussion difference in mean age between these two groups,
Our findings in 399 subjects are consistent indicating that the discrepancy between antibody and
with previous reports showing that the prevalence of staining could not be accounted for by age-related
H. pylori antibody increases with age and is associ- increases in H. pylori seroprevalence. Analysis of the
ated with gastritis (3-6,16,20-26). H. pylori antibody data in Table 1 by stepwise logistic regression also
correlated well with gastritis overall; only 3% of showed that age was not an important factor for
subjects with normal histology of the body and an- distinguishing seropositive and seronegative subjects.
trum were seropositive, whereas 99% of those with Gastritis and serological markers of gastritis were the
some degree of gastritis affecting either the body or important variables.
antrum were seropositive. We believe that the relatively high prevalence of H.
Among the 124 subjects who had gastric biopsy pylori antibody and low H. pylori staining peculiar to
specimens evaluated for H. pylori by Giemsa staining, subjects with atrophic body gastritis may reflect prior
antibody and staining results correlated well except or hidden infection with H. pylori in this subgroup.
in subjects with atrophic body gastritis. The preva- Ordinarily, the presence of serum antibodies to H.
pylori antigens as determined by ELISA correlates
well with measures of active colonization such as
Table 5. Comparison of Measured Parameters Between
staining, culture, and carbon isotope urea breath tests
Subjects With Positive and Negative Giemsa Stain
Results Among Subjects With Positive Serum
(14,24,28-29). The use of ELISA to detect prior infec-
Antibody to H. pylori tion with H. pylori depends on the persistence of
antibody after disappearance of the organism. Prelim-
Seropositive subjects (mean 2 SE)
inary reports suggest that H. pylori antibody concen-
Negative stain Positive stain tration in serum decreases with time after the organ-
n = 25 n = 59 P ism is eradicated with antibiotics and bismuth-
Sex (% male) 65 50 NS containing compounds (30,3 1). The length of time
Age (yr) 57.3 (3.4) 57.2 (1.8) NS required for H. pylori antibody concentration to de-
PGI (ng/mL) 41.1 (7.4) 80.9 (4.8) <0.0001 crease below defined seropositive cutoffs is not known,
PGII (ngimL) 22.5 (2.2) 22.9 (1.4) NS
PGI-PGII ratio 1.7 (0.2) 3.8 (0.2) <0.0001 but, in the absence of evidence of colonization, the
Gastrin (pg/mL) 99.3 (18.5) 63.3 (8.0) (0.05 presence of antibody to H. pylori should reflect previ-
MAO (mmollh) 5.4 (2.1) 21.6 (1.9) <0.0001 ous or current hidden infection.
Antral grade (l-7) 3.5 (0.4) 4.3 (0.2) < 0.05 An important question raised by our findings con-
Body grade (l-7) 5.4 (0.3) 3.4 (0.2) <0.0001
cerns the possible role of chronic H. pylori infection in
172 KARNES ET AL. GASTROENTEROLOGY Vol. 101, No. 1

4 5 6

Grade of body ga.Wftis MAO (mmoVh) Gastrin (pglml)

ii
Figure 3. Prevalence of H. py-
lori seropositivity (open cir-
2% 26 18 21 (M cles] and positive H. pylori
stains (closed circles) accord-

I 1P ing to histological grade of

body gastritis (A), histologi-
cal grade of antral gastritis
(B), MAO (C), serum PGI level
(D), serum gastrin concentra-
‘la0 tion (IS)and age (F). *Signifi-
5 6
cant differences by 2 analy-
Grade of Antral Gastritis PG I (@ml) sis (P < 0.05).

the pathogenesis of atrophic body gastritis. Before
the discovery of H. pylori, Ihamaki et al. (8) found
that in 42% of Finnish subjects with superficial
gastritis atrophic gastritis developed over three de-
cades. Specifically, they found that antral mucosa
often improved histologically, whereas superficial
body gastritis often progressed toward atrophy. These
observations, which were corroborated by the same
investigators in a cross-sectional study (32), indicate
that superficial gastritis may advance over 20-30
years into atrophic body gastritis. It is tempting to
speculate that the cases of superficial gastritis that
progressed to atophic body gastritis were caused by H.
pylori.
The hypothesis that atrophic body gastritis repre-
sents an end stage of H. pylori-associated superficial
gastritis has not attracted much attention because of
the relatively low prevalence of gastric colonization
with H. pylori in humans with atrophic body gastritis
(3) or pernicious anemia (33-35). As proposed by
DeLuca (36), this hypothesis requires that H. pylori
initiate an irreversible process toward atrophic gastri-
tis accompanied by a dramatic reduction or even
demise of the organism. Disappearance of the organ-
ism would correlate with the development of intesti-
nal metaplasia and hypochlorhydria, conditions that
seem to be inhospitable for H. pylori (3,37). Our data
are consistent with this hypothesis. However, none of
our subjects with atrophic body gastritis had abnor-
mal vitamin B,, levels or antibodies to intrinsic factor,
indicating that DeLuca’s hypothesis may be true for a
subgroup of individuals with atrophic body gastritis
who do not have pernicious anemia.
An alternative explanation for the high prevalence
of positive H. pylori antibody in the subgroup with
atrophic body gastritis with negative staining is in-

Table 6. Proportion of Subjects With Positive Serum Antibody and Positive Tissue Staining for H. pylori As a Function of
MucosaJ Histology

Body

Normal Superficial Atrophic

Antibody Stain Antibody Stain Antibody Stain

Antrum
Normal O/28 l/28 O/l O/l 819” 219
(0%) (4%) (89%) (22%)
Superficial O/l 111 26129 26129 lO/ll” 4/l 1
(90%) (90%) (91%) (36%)
Atrophic l/2 212 26127 25127 13/16’ 6116
(96%) (93%) (81%) (38%)
Total l/31 4131 52156 51156 31/36’ 12/36
(3%) (13%) (93%) (91%) (86%) (33%)

“Proportion with positive serum antibody is significantly greater than proportion with positive tissue staining in subjects with atrophic body
gastritis (P < 0.05).
July 1991 H. FYLORIANTIBODY AND STAINING IN ATROPHIC BODY GASTRITIS
( r = -.02 (All 399 subjects)
i References
3
H. pylori ELBA
(Absorbance at 492 nm)
Figure 4. Scatter plot of C. jejuni ELISA and H. pyZori ELISA
absorbance readings at 492 MI. Small open circles represent all
399 subjects. No significant correlation was seen between C.
jejuni and H. pylori ELISA absorbance readings (r = -0.02).
Large open circles indicate subjects who had positive H. pylori
stain and positive H. pylori ELISA results. Closed circles indicate
subjects who had negative H. pylori stain and positive IX pylori
ELISA results. No significance difference was seen between these
two groups with respect to C. jejuni ELISA absorbance results.
creased exposure to antigenically cross-reacting en-
teric organisms secondary to gastric hypochlorhydria.
To explore this possibility, we screened all sera for
antibody to C. jeiuni, a bacteria that shares some
antigens with H. pylori (16-19).Recently, Perez-Perez
et al. showed that their H. pylori antigen preparation
did not share cross-reactivity with C. jejuni antigens
(38). Using our Zf. pylori antigen preparation, we also
found no evidence for cross-reactivity with a crude
preparation of C. jejuni antigens (r = 0.02). In particu-
lar, subjects with false positive H. pylori antibody did
not have a significantly higher mean C. jejuni ELISA
absorbance than subjects with true positive H, pylori
antibody.
Although our data are consistent with the hypothe-
sis that H. pylori may cause atrophic body gastritis,
proof of this hypothesis would require a longitudinal
study in which middle-aged subjects with H. pylori-
associated superficial gastritis are randomly assigned
to have their infections eradicated. Repeated biopsies
of the gastric mucosa over a several year follow-up
would indicate whether eradication of J-f. pylori re-
duced or prevented the development of atrophic body
gastritis.
In summary, we have shown that subjects with
atrophic body gastritis, with or without accompany-
ing antral gastritis, have a high prevalence of H. pylori
seropositivity and a low prevalence of positive H.
pylori staining. This observation suggests that most
patients with atrophic body gastritis have been in-
fected with H. pylori at some point in their life. The
marked discrepancy between the prevalence of posi-
173
tive H. pylori serology and staining in these subjects is
consistent with the hypothesis that H. pylori may
initiate an irreversible process toward atrophic gastri-
tis that ultimately leads to its demise (36). Proof of this
hypothesis awaits the results of longitudinal studies
comparing gastric histologies in subjects whose H.
pylori infections have been eradicated with those
whose H. pylori infections have been allowed to
continue for several years.
1. Graham DY, Adam E, Klein PD, Evans DJ Jr, Evans DG, Hazel1
SL, Alpert LC, Michaletz PA, Yoshimura HH. Epidemiology of
Campylobacfer pylori infection. Gastroenterol Clin Biol 1989;
13:84B-88B.
2. Goodwin CS, Armstrong JA, Marshall BJ. Campylobacterpylori-
dis, gastritis, and peptic ulceration. J Clin Path01 1986;39:353-
365.
3. Siurala M, Sipponen P, Kekki M. Campylobacter pylori in a
sample of Finnish population: relations to morphology and
functions of the gastric mucosa. Gut 1988;29:909-915.
4. Blaser MJ. Type B gastritis, aging, and Campylobacter pylori.
Arch Intern Med 1988;148:1021-1022.
5. Graham DY, Klein PD. Compylobacter pyloridis gastritis: the
past, the present, and speculations about the future. Am J
Gastroenterol 1987;82:283-286.
6. Blaser MJ. Gastric Campylobacter-like organisms, gastritis, and
peptic ulcer disease. Gastroenterology 1987;93:371-383.
7. Gutierrez 0, Sierra F, Gomez MC, Camargo H. Campylobacter
pylori in chronic environmental gastritis and duodenal ulcer
patients (abstr). Gastroenterology 1989;94:A163.
8. Ihamaki T, Kekki M, Sipponen P, Siurala M. The sequelae and
course of chronic gastritis during a 30- to 34-year bioptic
follow-up study. Stand J Gastroenterol 1985;20:485-491.
9. Ihamaki T, Varis K, Siurala M. Morphological, functional and
immunological state of the gastric mucosa in gastric carcinoma
families. Stand J Gastroenterol 1979;14:801-812.
10. Varis K, Stenman UH, Lehtola J, Siurala M. Gastric lesion and
pernicious anemia: a family study. Acta Hepatogastroenterol
(Stuttg) 1978;25:62-67.
11. Samloff IM, Varis K, Ihamaki T, Siurala M, Rotter JI. Relation-
ships among serum pepsinogen I, serum pepsinogen II, and
gastric mucosal histology. Gastroenterology 1982;83:204-209.
12. Kekki M, Siurala M, Varis K, Sipponen P, Sistonen P, Nevan-
linna H. Classification principles and genetics of chronic
gastritis. Stand J Gastroenterol 1987;141 (Suppl):22:1-12,
13. Grey SF, Wyatt JI, Rathbone BJ. Simplified techniques for
identifying gastric Campylobacterpyloridis on tissue sections. J
Clin Path01 1986;39:1279-1280.
14. Evans DJ Jr, Evans DG, Graham DY, Klein PD. A sensitive and
specific serologic test for detection of Compylobacter pylori
infection. Gastroenterology 1989;96:1004-1008.
35. Karnes WE Jr, Ohning GV, Sytnik B, Kim SWR, Walsh JH.
Preservation of pH inhibition of gastrin release in subjects with
Helicobacterpylori-associated hypergastrinemia. Rev Infect Dis
1990 (in press).
16. Jones DM, Lessells AM, Eldridge J. Campylobacter like organ-
isms on the gastric mucosa: culture, hisyological, and serologi-
cal studies. J Clin Path01 1984;37:1002-1006.
17. von Wulffen H, Heesemann J, Butzow GH, Loning T, Laufs R.
Detection of Compylobacter pyloridis in patients with antrum
gastritis and peptic ulcers by culture, complement fixation test,
and immunoblot. J Clin Microbial 1986;24:716-720.
18. Jones DM, Eldridge J, Fox AJ, Sethi P, Whorwell PJ, Antibody to
174 KARNES ET AL.
the gastric campylobacter-like organism (“Campylobacter
pyZoridis”)-clinical correlations and distribution in the nor-
mal population. J Med Microbial 1986;22:57-62.
19. Kaldor J, Tee W, Nicolacopolous C, Demirtzoglou K, Noonan D,
Dwyer B. Immunoblot confirmation of immune response to
Campylobacterpyloridis in patients with duodenal ulcers. Med
JAust 1986;145:133-135.
20. Blomberg B, Jarnerot G, Kjellander J, Danielsson D, Kraaz W.
Prevalence of Compylobacter pylori in an unselected Swedish
population of patients referred for gastroscopy. Stand J Gastro-
enter01 1988;23:358-362.
21. Hazel1 SL, Hennessy WB, Borody TJ, Carrick J, Ralston M,
Brady L, Lee A. Campylobacter pyloridis gastritis II: distribu-
tion of bacteria and associated inflammation in the gastroduo-
denal environment. Am J Gastroenterol1987;82:297-301.
22. Lamouliatte H, Megraud F, De Mascarel A, Roux D, Quinton A.
“Campylobacter pyloridis” and epigastric pain: endoscopic,
histological, and bacteriological correlations. Gastroenterol
Clin Biol 1987;11:212-216.
23. Yardley JH, Paul1 G. Campylobacter pylori: a newly recognized
infectious agent in the gastrointestinal tract. Am J Surg Path01
1988;12:89-99.
24. Perez Perez GI, Dworkin BM, Chodos JE, Blaser MJ. Campylo-
bacterpylori antibodies in humans. Ann Intern Med 1988;109:
11-17.
25. Morris A, Nicholson G, Lloyd G, Haines D, Rogers A, Taylor D.
Seroepidemiology of Campylobacter pyloridis. N Z Med J
1986;99:657-659.
26. Booth L, Holdstock G, MacBride H, Hawtin P, Gibson JR,
Ireland A, Bamforth J, DuBoulay CE, Lloyd RS, Pearson AD.
Clinical importance of Campylobacterpyloridis and associated
serum IgG and IgA antibody responses in patients undergoing
upper gastrointestinal endoscopy. J Clin Path01 1986;39:215-
219.
27. Fox JG, Correa P, Taylor NS, Zavala D, Fontham E, Janney F,
Rodriguez E, Hunter F, Diavolitsis S. Campylobocter pylori-
associated gastritis and immune response in a population at
increased risk of gastric carcinoma. Am J Gastroenterol1989;84:
775-781.
28. Goodwin CS, Blincow E, Peterson G, Sanderson C, Cheng W,
Marshall B, Warren JR, McCulloch R. Enzyme-linked immu-
nosorbent assay for Campylobacter pyloridis: correlation with
presence of C. pyloridis in the gastric mucosa. J Infect Dis
1987;155:488-494.
GASTROENTEROLOGY Vol. 101,No. 1
29. Bolton FJ, Hutchinson DN. Evaluation of three Campylobacter
pylori antigen preparations for screening sera from patients
undergoing endoscopy. J Clin Path01 1989;42:723-726.
30. Vaira D, Holton J, Cairns SR, Falzon M, Polydorou A, Dowsett
JF, Salmon PR. Antibody titres to Campylobacter pylori after
treatment for gastritis. BMJ 1988;297:397-397.
31. van Bohemen CG, Langenberg ML, Rauws EA, Oudbier J,
Weterings E, Zanen HC. Rapidly decreased serum IgG to
Campylobacter pylori following elimination of Campylobacter
in histological chronic biopsy Campylobacter-positive gastri-
tis. Immunol Lett 1989;20:59-61.
32. Kekki M, Varis K, Pohjanpalo H, Isokoski M, Ihamaki T, Siurala
M. Course of antrum and body gastritis in pernicious anemia
families. Dig Dis Sci 1983;28:698-704.
33. Flejou JF, Bahame P, Smith AC, Stockbrugger RW, Rhode J,
Price AB. Pernicious anemia and campylobacter-like organ-
isms; is the gastric antrum resistant to colonization? Gut
1989;30:60-64.
34. OConnor HJ, Axon ATR, Doxon MF. Campylobacter-like organ-
isms unusual in type A (pernicious anemia) gastritis. Lancet
1984;1:57.
35. Gonzalez JD, Sancho FJ, Sainz S, Such J, Fernandez M, Mones
Xiol J. Campylobacter pylori and pernicious anemia (letter).
Lancet 1988;1(8575-8576):57.
36. DeLuca VA Jr. Is pernicious anemia caused by Campylobacter
pylori gastritis [letter]. J Clin Gastroenterol 1989;11:584-585.
37. Fiocca R, Villani L, Turpini F, Turpini R, Solcia E. High
incidence of Campylobacter-like organisms in endoscopic biop-
sies from patients with gastritis, with or without peptic ulcer.
Digestion 1987;38:234-244.
38. Perez-Perez GI, Taylor DN, Bodhidatta L, Wongsrichanalai J,
Baze WB, Dunn BE, Echeverria PD, Blaser MJ. Seroprevalence
of Helicobacfer pylori infections in Thailand. J Infect Dis
1990;161:1237-1271.
Received July 5,199O.Accepted December 9,199O.
Address requests for reprints to: John H. Walsh, M.D., CURE VA
Wodsworth, Room 115, Building 115,Wilshire and Sawtelle Boule-
vard, Los Angeles, California 90973.
This work was supported by NIH grant DK17328 and by a grant
from Procter&Gamble Co., Cincinnati, Ohio.